YY1-induced lncRNA ZFPM2-AS1 facilitates cell proliferation and invasion in small cell lung cancer via upregulating of TRAF4

Background Newly identified lncRNA zinc finger protein, FOG family member 2 antisense RNA 1 (ZFPM2-AS1) is identified as an oncogenic gene. However, the role of ZFPM2-AS1 in small cell lung cancer (SCLC) is poorly comprehended. Methods The expression of genes in SCLC tissues and cells was measured by qRT-PCR. Colony formation, EdU, CCK-8, transwell and wound healing as well as in vivo assays revealed the function of ZFPM2-AS1 in SCLC. ChIP, luciferase reporter, RIP and RNA pull down assays demonstrated the binding relation among genes. Results ZFPM2-AS1 was significantly upregulated in SCLC tissues and cells. ZFPM2-AS1 deficiency attenuated SCLC cell proliferation, invasion and migration. In addition, ZFPM2-AS1 was transcriptionally activated by Yin Yang 1 (YY1) factor. Further, miR-3612 was confirmed as downstream miRNA of ZFPM2-AS1. Moreover, TNF receptor associated factor 4 (TRAF4) was the target gene of miR-3612 in SCLC. ZFPM2-AS1, miR-3612 and TRAF4 jointly constituted a competing endogenous RNA (ceRNA) network in SCLC. Finally, TRAF4 could countervail ZFPM2-AS1 downregulation-mediated function on SCLC cell proliferation and invasion in vitro and tumor growth in vivo. Conclusion Our study elucidated the oncogenic effect of ZFPM2-AS1 in SCLC progression, indicating it may be a therapeutic target for SCLC.


Background
Statistically, lung cancer is characterized with the highest incidence, occupying 11.6% of cancer-induced mortalities [1]. Domestically, the estimated number of diagnosed lung cancer was 733,300 with 610,200 death cases in 2015, indicating the life-threatening influence of lung cancer [2]. Compared with non-small-cell lung cancer (NSCLC), small-cell lung cancer (SCLC) only accounts for a minor part in lung cancer. More surprisingly, SCLC is substantiated as the subtype with highest risk of aggressiveness due to its rapid growth and early invasiveness [3,4]. Additionally, 5-year survival of SCLC is lower than 10%, meanwhile, the dreadful median survival of it is unsatisfactory [5][6][7]. Nevertheless, the number of existing research focusing on SCLC is limited.
The transcribed RNA contains non-coding RNAs (ncR-NAs) in large quantity, which are limited in protein-coding [8]. Long non-coding ncRNAs (lncRNAs), containing over 200 nucleotides in length, belong to ncRNAs. Further lncRNAs are validated to exert multiple functions in biological processes [9]. Their epigenetic silencing was constituted by lncRNA-miRNA interplay and lncRNAprotein interaction [10]. MicroRNAs (miRNAs), typically 18-200 nucleotides in length, has extensively reported by researches in tumor, and up to 30% of mRNAs are regulated by thousands of miRNAs [9]. LncRNAs could competitively bind miRNAs to modulate the expression of target mRNAs, forming a ceRNA network [11,12]. For instance, lncRNA PVT1 enhances insulin like growth factor 1 receptor (IGF1R) expression to promote cell proliferation and invasion in papillary thyroid carcinoma by acting as ceRNA of miRNA-30a [13]. LncRNA RP4 serves as a ceRNA for SH3GLB1 to sponge miR-7-5p in colorectal cancer [14]. LINC00488 upregulates TLN1 as a ceRNA in hepatocellular carcinoma cell growth through sponging miR-330-5 [15]. Above all, lncRNAs could affect tumorigenesis via gene modulation. Moreover, various cancers including SCLC are functionally associated with abnormally expressed lncRNAs [16,17]. For example, highly expressed AFAP1-AS1 is correlated with nasopharyngeal carcinoma metastasis [18]. And overexpressed LINC00511 is implicated in breast cancer tumourigenesis and stemness [19]. Nevertheless, the investigation about participation of lncRNAs in SCLC is limited.

Patient tissues
In this study, total 54 pairs of small cell lung cancer (SCLC) tissues and matched normal tissues were collected under the approval from the Ethics Committee of Minhang Hospital, Fudan University. All participants were not treated with radiotherapy and chemotherapy before operation. Tumor tissues and respective normal tissues were snap-frozen in liquid nitrogen at − 80 °C until requested.

Quantitative real-time polymerase chain reaction (qRT-PCR)
TRIzol reagent (Invitrogen, Carlsbad, CA, USA) was applied to extract total RNA from cells. Then, Reverse Transcription Kit was used to conduct the reverse transcription between total RNA and cDNA. qRT-PCR was performed in line with SYBR-Green Real-Time PCR Kit (Takara, Tokyo, Japan). Relative gene expression was operated with 2 −ΔΔCt method and normalized to the GAPDH or U6.

Colony formation assay
DMS-53 or SHP-77 cells were cultivated in 6-well plates after transfection. Two weeks later, cells were washed with PBS (Solibao technology, Beijing, China), fixed in methanol (Solibao technology) and stained with crystal violet (Solibao technology). Finally, the colony numbers were counted manually.

Cell proliferation assay
The viability of DMS-53 or SHP-77 cells was detected by the use of CCK-8 Kit (Sigma-Aldrich, Carlsbad, USA). Cells were placed into 96-well plates at specific time points. 10 μl CCK-8 were added into each well for 4 h. To measure the absorbance at 450 nm, the microplate reader (Bio-Rad, Hercules, CA, USA) was utilized.

Transwell assay
2 × 10 4 cells were re-suspended in serum-free medium and were cultured in the top transwell chamber (BD Biosciences, NY, USA). Then, 10% PBS was added to the lower chamber. Migrated cells were fixed with methanol and dyed in crystal violet, followed by observation under microscope (Leica Microsystems, Wetzlar, Germany). Invasion assay was conducted the same as above, but pre-coated with matrigel (BD Bioscience).

EdU incorporation assay
Based upon the standardized guides, cells were cultured with DMEM supplemented with 300 μl EdU for 2 h after transfection. Then, cells were fixed with 4% formaldehyde (Solibao technology) at room temperature for 30 min. After staining, cells were observed under fluorescence microscope (Leica Microsystems). Besides, EdU-positive cells (stained in red) are cells that were proliferating, and DAPI stained cells (stained in blue) were total cells.

Wound healing assay
Cells were seeded to 6-well plates and cultivated with general cell growth fluid. The sterile pipette was used to make scratches in cell layers when cells were paved. Once cells were cleaned, the medium was replaced. Ultimately, the images were collected and analyzed.

Western blot
The cells were lysed with RIPA lysis buffer containing protease inhibitor and the complete DNA was obtained. After detecting the protein concentrations by BCA method, proteins were isolated by sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) gel electrophoresis, and were then transferred to PVDF. The membrane was incubated with primary antibodies overnight after PVDF was sealed with skim milk. The primary antibodies including: anti-YY1 (ab109237, Abcam, Cambridge, USA), anti-CBX5 (ab109028, Abcam) anti-TRAF4 (ab190986, Abcam), anti-TAOK1 (ab197891, Abcam) and anti-GAPDH (ab8245, Abcam) at 4 °C. Following that, the chemiluminescence system (Hubei Biossci Biotechnology, Wuhan, China) was employed for detecting the protein.

Chromatin immunoprecipitation (ChIP) assay
ChIP assay was implemented with the application of EZ ChIP ™ Chromatin Immunoprecipitation Kit for cells (Millipore, Bedford, MA), following then standard method. DMS-53 or SHP-77 cells were fixed with 1% formaldehyde for 15 min's crosslink and lysed, and then the DNA was sonicated for shearing DNA into 500-bp fragments. DNA samples were precipitated with anti-IgG or anti-YY1 antibody for 6 h, in the presence of 30 μl of magnetic beads. Subsequently, the precipitated chromatin DNA was collected, extracted and purified for qRT-PCR analysis.

Subcellular fractionation
According to the indicated protocols, Cytoplasmic and Nuclear RNA Purification Kit (Norgen, Ontario, Canada) was employed for isolating and purifying cytoplasmic and nuclear RNA. Expression levels of ZFPM2-AS1 were assessed by qRT-PCR analysis. GAPDH and U1 were used to control of cytoplasm or nucleus.

RNA pull down assay
Bio-miR-miR-3612-Mut/Wt and Bio-NC were acquired via biotin-labeling miR-miR-3612-Wt/Mut and NC. ZFPM2-AS1 biotin probe and no ZFPM2-AS1 probe constructed by GenePharma were treated with M-280 Streptavidin magnetic beads (Invitrogen). Then, cells were collected, lysed, and incubated with probe-coated beads overnight for 48 h. RNA complexes after purification were detected by qRT-PCR assay.

RNA immunoprecipitation (RIP) assay
Cells were lysed, incubated with anti-Ago2 and anti-IgG antibodies (Millipore, Bellerica, MA, USA) coated on magnetic beads in RIP buffer. The precipitated RNAs were isolated for qRT-PCR.

Fluorescence in situ hybridization (FISH) assay
ZFPM2-AS1 subcellular localization was detected using a FISH Kit (Roche, Basel, Switzerland). Cells were mixed with paraformaldehyde for fixation. Then, ZFPM2-AS1 probe (Sigma-Aldrich) hybridization solution was added. Cell nucleus was stained for 10 min at room temperature. Lastly, the laser confocal scanning microscopy (FV1000; Olympus, Tokyo, Japan) was used.

Animal study
Female BALB/C athymic nude mice (18-20 g; 6-weekold) were procured from the National Laboratory Animal Center (Beijing, China) and kept under the SPF-grade animal lab. The animal study was undertaken with the approval from the Animal Research Ethics Committee of Minhang Hospital, Fudan University. In vivo study was implemented via subcutaneous injection of 5 × 10 6 transfected SCLC cells to mice for 28 days, with tumor volume monitored every 4 days. The tumors collected from killed mice were weighed for analysis.

Statistical analysis
GraphPad Prism 7 software package (Graph-Pad Software, Inc., La Jolla, CA, USA) was used for the statistical analysis. Quantitative data were showed as mean ± standard deviation. Student's-test and ANOVA was used to compare the difference of groups. P < 0.05 had statistically significance. The experiment was conducted at least in thrice.

ZFPM2-AS1 upregulated TRAF4 expression via sequestering miR-3612
The decisive role of mRNAs in the regulating cancer progression induced us to find out the target gene of miR-3612. RNA22, microT, PicTar, and miRmap were employed for the exploration of miR-3612 targets. Three genes (CBX5, TRAF4, TAOK1) were revealed as candidates (Fig. 4a). Via the detections of qRT-PCR and western blot, miR-3612 negatively regulated TRAF4 expression in SHP-77 and DMS-53 cells (Fig. 4b). Additionally, TRAF4 was highly expressed in SCLC tissues rather normal tissues (Additional file 1: Fig. S1E). In a similar way, the predicted binding site of TRAF4 and miR-3612 was obtained. Luciferase reporter assay was conducted to evaluate the effectiveness of the indicated binding site. The activity of TRAF4-WT in SHP-77 and DMS-53 cells was repressed by miR-3612 mimics or simulated by its inhibitor (Fig. 4c, d). Besides, RIP assay demonstrated the co-existence of ZFPM2-AS1, miR-3612 and TRAF4 in Ago2-specific complex (Fig. 4e). Also, the enrichment of ZFPM2-AS1 and TRAF4 was detected in Bio-miR-3612 sense group (Fig. 4f ). Subsequently, it verified that silencing of ZFPM2-AS1 decreased TRAF4 mRNA and protein expressions, while suppressing miR-3612 could countervail this repressing effect (Fig. 4g, h). It concluded that ZFPM2-AS1 competitively bound miR-3612 to upregulate TRAF4 expression in SCLC cells.

ZFPM2-AS1 contributed to cell proliferation, invasion, migration and tumor growth in SCLC via upregulating TRAF4
To certify whether TRAF4 affected the modulation function of ZFPM2-AS1 in SCLC cell growth, several rescue experiments were carried out. TRAF4 was overexpressed firstly, and its mRNA and protein levels were enhanced remarkably (Fig. 5a). Afterwards, TRAF4 elevation abolished the hindering effects of ZFPM2-AS1 depletion on cell proliferation (Fig. 5b-d). ZFPM2-AS1 inhibition induced the suppressed cell migration and invasion, while upregulating TRAF4 expression counteracted this inhibitory influence (Fig. 5e). Wound healing assay revealed that overexpression of TRAF4 attenuated the suppressive role of ZFPM2-AS1 deficiency on cell migration (Fig. 5f, Additional file 2: Figure S2A). On the other hand, in vivo experiments were also performed. The tumors extracted from mice injected with sh/ZFPM2-AS1#2-transfected cells were smaller and lighter than NC group. Then TRAF4 overexpression offset the suppressing effects of ZFPM2-AS1 depletion on tumor size, volume and weight (Additional file 2: Fig. S2B-D). In

Discussion
Recently, the biological importance of multiple lncR-NAs in tumor progression has been emphasized extensively, indicating the possibility of the molecular-targeted application in treatment of tumors including SCLC. And the key role of lncRNAs in SCLC progression has been reported. For example, lncRNA HOTAIR is associated with cellular proliferation, invasion, and clinical reoccurrence in SCLC [25]. LncRNA TUG1 affects SCLC cell growth via regulating LIMK2b through EZH2 [26]. LncRNA PVT1 overexpression serves as a poor prognostic maker in SCLC, and facilitates malignant migration and invasion of cells [27]. In our research, lncRNA ZFPM2-AS1, located in 8q23.1, was an oncogenic gene in gastric cancer, renal cell cancer and lung adenocarcinoma [20,22,23]. Nonetheless, the role of ZFPM2-AS1 in SCLC was rarely discussed. In our research, we found its significantly high expression tendency in SCLC tissues and cell lines. Functionally, ZFPM2-AS1 insufficiency inhibited SCLC cell proliferation, invasion and migration in vitro and tumor growth in vivo, suggesting the carcinogenic role of ZFPM2-AS1 in SCLC. Accumulating literatures showed that lncRNAs transcription could be attributed to specific transcriptional factors, including YY1. For instance, c-Myc activates lncRNA CCAT1 expression in gastric cancer [28]. STAT3 regulates the upregulation of lncRNA to affect in liver cancer metastasis [29]. YY1 contributes to lncRNA-PVT1 activation during lung cancer progression [30]. Besides, YY1, a ubiquitous protein in normal and tumor tissues, is identified as a potential prognosis biomarker and therapeutic target [31]. In this study, YY1 could bind with ZFPM2-AS1 promoter and transcriptionally activated ZFPM2-AS1 in SCLC cells.
After comprehending upstream regulation mechanism of ZFPM2-AS1, the downstream regulation mechanism also deserved to be discussed. Cytoplasmic lncRNAs participate in the ceRNA pattern in tumors. For illustration, lncRNA HOTAIR promotes HER2 expression through sponging miR-331-3p in gastric cancer [32]. LncRNA Unigene56159 accelerates EMT process in hepatocellular carcinoma via sponging miR-140-5p to regulate Slug [33]. Concerning the mechanism underlying ZFPM2-AS1 in SCLC, we hypothesized the ceRNA role of ZFPM2-AS1 in SCLC due to its large portion in SCLC cytoplasm. As expected, ZFPM2-AS1 could sponge miR-3612 to release TRAF4 expression. Moreover, the tumor-facilitator role of TRAF4 has been validated in hepatocellular carcinoma [34], intrahepatic cholangiocarcinoma [35], breast cancer [36], and colon cancer [37]. More importantly, TRAF4 is confirmed to promote the development of lung cancer [38][39][40]. And our study first illuminated that TRAF4 was involved in SCLC progression. TRAF4 overexpression reversed the suppressive function of ZFPM2-AS1 depletion on SCLC cell proliferation, invasion and migration in vitro as well as tumor growth in vivo.

Conclusion
YY1-activated ZFPM2-AS1 promoted the malignant phenotypes of SCLC via sequestering miR-3612 to upregulate TRAF4, which possibly provide a promising candidate target for SCLC treatment.
Additional file 1: Figure S1. (A) The expression of ZFPM2-AS1 in SCLC tissues and matched normal tissues was detected by qRT-PCR. (B) The original picture of wound healing assay in Fig. 1g. (C-E) qRT-PCR measured the expression of YY1, miR-3612 and TRAF4 in SCLC tissues and matched normal tissues. **P < 0.01 Additional file 2: Figure S2. (A) The original picture of wound healing assay in Fig. 5f. (B-D) The pictures of tumors obtained from mice injected with differently transfected cells were taken. Tumor volume and weight were also measured. **P < 0.01